"Because chatbots don’t have a body or internal state to manage, they don’t have “regulatory objectives.” Without the physical limits of a biological body, AI models churn out data without caution, leading to unsafe, overconfident, and untrustworthy answers."